| Kconfig (902861e34c401696ed9ad17a54c8790e7e8e3069) | Kconfig (cd14b01846612f3f3277e97bfbecba4c8cee5ce9) |
|---|---|
| 1# SPDX-License-Identifier: GPL-2.0-only 2 3menu "Memory Management options" 4 5# 6# For some reason microblaze and nios2 hard code SWAP=n. Hopefully we can 7# add proper SWAP support to them, in which case this can be remove. 8# --- 31 unchanged lines hidden (view full) --- 40 depends on ZSWAP 41 help 42 If selected, the compressed cache for swap pages will be enabled 43 at boot, otherwise it will be disabled. 44 45 The selection made here can be overridden by using the kernel 46 command line 'zswap.enabled=' option. 47 | 1# SPDX-License-Identifier: GPL-2.0-only 2 3menu "Memory Management options" 4 5# 6# For some reason microblaze and nios2 hard code SWAP=n. Hopefully we can 7# add proper SWAP support to them, in which case this can be remove. 8# --- 31 unchanged lines hidden (view full) --- 40 depends on ZSWAP 41 help 42 If selected, the compressed cache for swap pages will be enabled 43 at boot, otherwise it will be disabled. 44 45 The selection made here can be overridden by using the kernel 46 command line 'zswap.enabled=' option. 47 |
| 48config ZSWAP_EXCLUSIVE_LOADS_DEFAULT_ON 49 bool "Invalidate zswap entries when pages are loaded" 50 depends on ZSWAP 51 help 52 If selected, exclusive loads for zswap will be enabled at boot, 53 otherwise it will be disabled. 54 55 If exclusive loads are enabled, when a page is loaded from zswap, 56 the zswap entry is invalidated at once, as opposed to leaving it 57 in zswap until the swap entry is freed. 58 59 This avoids having two copies of the same page in memory 60 (compressed and uncompressed) after faulting in a page from zswap. 61 The cost is that if the page was never dirtied and needs to be 62 swapped out again, it will be re-compressed. 63 |
|
| 48config ZSWAP_SHRINKER_DEFAULT_ON 49 bool "Shrink the zswap pool on memory pressure" 50 depends on ZSWAP 51 default n 52 help 53 If selected, the zswap shrinker will be enabled, and the pages 54 stored in the zswap pool will become available for reclaim (i.e 55 written back to the backing swap device) on memory pressure. --- 522 unchanged lines hidden (view full) --- 578# support for memory balloon 579config MEMORY_BALLOON 580 bool 581 582# 583# support for memory balloon compaction 584config BALLOON_COMPACTION 585 bool "Allow for balloon memory compaction/migration" | 64config ZSWAP_SHRINKER_DEFAULT_ON 65 bool "Shrink the zswap pool on memory pressure" 66 depends on ZSWAP 67 default n 68 help 69 If selected, the zswap shrinker will be enabled, and the pages 70 stored in the zswap pool will become available for reclaim (i.e 71 written back to the backing swap device) on memory pressure. --- 522 unchanged lines hidden (view full) --- 594# support for memory balloon 595config MEMORY_BALLOON 596 bool 597 598# 599# support for memory balloon compaction 600config BALLOON_COMPACTION 601 bool "Allow for balloon memory compaction/migration" |
| 586 def_bool y | 602 default y |
| 587 depends on COMPACTION && MEMORY_BALLOON 588 help 589 Memory fragmentation introduced by ballooning might reduce 590 significantly the number of 2MB contiguous memory blocks that can be 591 used within a guest, thus imposing performance penalties associated 592 with the reduced number of transparent huge pages that could be used 593 by the guest workload. Allowing the compaction & migration for memory 594 pages enlisted as being part of memory balloon devices avoids the 595 scenario aforementioned and helps improving memory defragmentation. 596 597# 598# support for memory compaction 599config COMPACTION 600 bool "Allow for memory compaction" | 603 depends on COMPACTION && MEMORY_BALLOON 604 help 605 Memory fragmentation introduced by ballooning might reduce 606 significantly the number of 2MB contiguous memory blocks that can be 607 used within a guest, thus imposing performance penalties associated 608 with the reduced number of transparent huge pages that could be used 609 by the guest workload. Allowing the compaction & migration for memory 610 pages enlisted as being part of memory balloon devices avoids the 611 scenario aforementioned and helps improving memory defragmentation. 612 613# 614# support for memory compaction 615config COMPACTION 616 bool "Allow for memory compaction" |
| 601 def_bool y | 617 default y |
| 602 select MIGRATION 603 depends on MMU 604 help 605 Compaction is the only memory management component to form 606 high order (larger physically contiguous) memory blocks 607 reliably. The page allocator relies on compaction heavily and 608 the lack of the feature can lead to unexpected OOM killer 609 invocations for high order memory requests. You shouldn't --- 6 unchanged lines hidden (view full) --- 616 depends on COMPACTION 617 default 0 if PREEMPT_RT 618 default 1 619 620# 621# support for free page reporting 622config PAGE_REPORTING 623 bool "Free page reporting" | 618 select MIGRATION 619 depends on MMU 620 help 621 Compaction is the only memory management component to form 622 high order (larger physically contiguous) memory blocks 623 reliably. The page allocator relies on compaction heavily and 624 the lack of the feature can lead to unexpected OOM killer 625 invocations for high order memory requests. You shouldn't --- 6 unchanged lines hidden (view full) --- 632 depends on COMPACTION 633 default 0 if PREEMPT_RT 634 default 1 635 636# 637# support for free page reporting 638config PAGE_REPORTING 639 bool "Free page reporting" |
| 624 def_bool n | |
| 625 help 626 Free page reporting allows for the incremental acquisition of 627 free pages from the buddy allocator for the purpose of reporting 628 those pages to another entity, such as a hypervisor, so that the 629 memory can be freed within the host for other uses. 630 631# 632# support for page migration 633# 634config MIGRATION 635 bool "Page migration" | 640 help 641 Free page reporting allows for the incremental acquisition of 642 free pages from the buddy allocator for the purpose of reporting 643 those pages to another entity, such as a hypervisor, so that the 644 memory can be freed within the host for other uses. 645 646# 647# support for page migration 648# 649config MIGRATION 650 bool "Page migration" |
| 636 def_bool y | 651 default y |
| 637 depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU 638 help 639 Allows the migration of the physical location of pages of processes 640 while the virtual addresses are not changed. This is useful in 641 two situations. The first is on NUMA systems to put pages nearer 642 to the processors accessing. The second is when allocating huge 643 pages as migration can relocate pages to satisfy a huge page 644 allocation instead of reclaiming. --- 235 unchanged lines hidden (view full) --- 880 subsystems to allocate big physically-contiguous blocks of memory. 881 CMA reserves a region of memory and allows only movable pages to 882 be allocated from it. This way, the kernel can use the memory for 883 pagecache and when a subsystem requests for contiguous area, the 884 allocated pages are migrated away to serve the contiguous request. 885 886 If unsure, say "n". 887 | 652 depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU 653 help 654 Allows the migration of the physical location of pages of processes 655 while the virtual addresses are not changed. This is useful in 656 two situations. The first is on NUMA systems to put pages nearer 657 to the processors accessing. The second is when allocating huge 658 pages as migration can relocate pages to satisfy a huge page 659 allocation instead of reclaiming. --- 235 unchanged lines hidden (view full) --- 895 subsystems to allocate big physically-contiguous blocks of memory. 896 CMA reserves a region of memory and allows only movable pages to 897 be allocated from it. This way, the kernel can use the memory for 898 pagecache and when a subsystem requests for contiguous area, the 899 allocated pages are migrated away to serve the contiguous request. 900 901 If unsure, say "n". 902 |
| 903config CMA_DEBUG 904 bool "CMA debug messages (DEVELOPMENT)" 905 depends on DEBUG_KERNEL && CMA 906 help 907 Turns on debug messages in CMA. This produces KERN_DEBUG 908 messages for every CMA call as well as various messages while 909 processing calls such as dma_alloc_from_contiguous(). 910 This option does not affect warning and error messages. 911 |
|
| 888config CMA_DEBUGFS 889 bool "CMA debugfs interface" 890 depends on CMA && DEBUG_FS 891 help 892 Turns on the DebugFS interface for CMA. 893 894config CMA_SYSFS 895 bool "CMA information through sysfs interface" 896 depends on CMA && SYSFS 897 help 898 This option exposes some sysfs attributes to get information 899 from CMA. 900 901config CMA_AREAS 902 int "Maximum count of the CMA areas" 903 depends on CMA | 912config CMA_DEBUGFS 913 bool "CMA debugfs interface" 914 depends on CMA && DEBUG_FS 915 help 916 Turns on the DebugFS interface for CMA. 917 918config CMA_SYSFS 919 bool "CMA information through sysfs interface" 920 depends on CMA && SYSFS 921 help 922 This option exposes some sysfs attributes to get information 923 from CMA. 924 925config CMA_AREAS 926 int "Maximum count of the CMA areas" 927 depends on CMA |
| 904 default 20 if NUMA 905 default 8 | 928 default 19 if NUMA 929 default 7 |
| 906 help 907 CMA allows to create CMA areas for particular purpose, mainly, 908 used as device private area. This parameter sets the maximum 909 number of CMA area in the system. 910 | 930 help 931 CMA allows to create CMA areas for particular purpose, mainly, 932 used as device private area. This parameter sets the maximum 933 number of CMA area in the system. 934 |
| 911 If unsure, leave the default value "8" in UMA and "20" in NUMA. | 935 If unsure, leave the default value "7" in UMA and "19" in NUMA. |
| 912 913config MEM_SOFT_DIRTY 914 bool "Track memory changes" 915 depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS 916 select PROC_PAGE_MONITOR 917 help 918 This option enables memory changes tracking by introducing a 919 soft-dirty bit on pte-s. This bit it set when someone writes --- 48 unchanged lines hidden (view full) --- 968 This feature allows to estimate the amount of user pages that have 969 not been touched during a given period of time. This information can 970 be useful to tune memory cgroup limits and/or for job placement 971 within a compute cluster. 972 973 See Documentation/admin-guide/mm/idle_page_tracking.rst for 974 more details. 975 | 936 937config MEM_SOFT_DIRTY 938 bool "Track memory changes" 939 depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS 940 select PROC_PAGE_MONITOR 941 help 942 This option enables memory changes tracking by introducing a 943 soft-dirty bit on pte-s. This bit it set when someone writes --- 48 unchanged lines hidden (view full) --- 992 This feature allows to estimate the amount of user pages that have 993 not been touched during a given period of time. This information can 994 be useful to tune memory cgroup limits and/or for job placement 995 within a compute cluster. 996 997 See Documentation/admin-guide/mm/idle_page_tracking.rst for 998 more details. 999 |
| 976# Architectures which implement cpu_dcache_is_aliasing() to query 977# whether the data caches are aliased (VIVT or VIPT with dcache 978# aliasing) need to select this. 979config ARCH_HAS_CPU_CACHE_ALIASING 980 bool 981 | |
| 982config ARCH_HAS_CACHE_LINE_SIZE 983 bool 984 985config ARCH_HAS_CURRENT_STACK_POINTER 986 bool 987 help 988 In support of HARDENED_USERCOPY performing stack variable lifetime 989 checking, an architecture-agnostic way to find the stack pointer --- 258 unchanged lines hidden --- | 1000config ARCH_HAS_CACHE_LINE_SIZE 1001 bool 1002 1003config ARCH_HAS_CURRENT_STACK_POINTER 1004 bool 1005 help 1006 In support of HARDENED_USERCOPY performing stack variable lifetime 1007 checking, an architecture-agnostic way to find the stack pointer --- 258 unchanged lines hidden --- |